The title of this paper stems from the observation of a significant difference between sandstone and carbonate reservoirs in their global temperature distributions of EUR. Using a global database with the largest 814 sandstone and 332 carbonate reservoirs (believed to be approximately representative of presently-known conventionally recoverable global petroleum volumes), it is found that the average temperature of EUR occurrence is much lower for sandstone reservoirs (75ºC) than for carbonate reservoirs (91ºC). Closer inspection reveals that this difference is present in the 869 oil reservoirs in our database, but that no overall difference is apparent between the 204 sandstone and 73 carbonate reservoirs in which gas dominates over oil. Furthermore, it is mainly the 23 largest oil reservoirs (>4 BBOE) that account for most of the overall sandstone-carbonate difference: 14 sandstones versus three carbonates at 100ºC.

Five possible reasons for the above difference are considered. Expulsion of oil at higher temperatures in carbonate source rocks could have this effect, but no evidence is known that supports this hypothesis. Calcareous shales (presumably more frequent as seals to carbonate reservoirs) may undergo illitization at higher temperatures than non-calcareous shales, but this does not explain why the difference in the temperature ranges of maximum EUR occurrences are observed for oil but not for gas reservoirs. Greater protection of porosity from high-temperature cementation in carbonate oil reservoirs is supported by the observation that this phenomenon is present in at least seven of the ten largest carbonate reservoirs at >100ºC, but this explanation seems inconsistent with the relative paucity of similarly abundant large carbonate reservoirs also at lower temperatures. There may also exist some systematic differences, as yet unknown, in the overall petroleum systems between the larger sandstone and carbonate reservoirs. Finally, it is possible that the observed difference is simply coincidental (not related to any causal process). Whatever the cause, the empirical data are real and are worthy of further investigation.